- Email: [email protected]
Restriction enzyme fragment length polymorphisms of amplified herpes simplex virus type-1 DNA provide epidemiologic information
129
DIAGN MICROBIOLINFECTDIS 1993;17:129-133
VIROLOGY
Restriction Enzyme Fragment Length Polymorphisms of Amplified Herpes Simplex Virus Type-1 DNA Provide Epidemiologic Information Thomas H. Haugen, Beth Alden, Sandra Matthey, and Donald Nicholson
Human herpes simplex type 1 (HSV-1) DNA of isolates from patients in a large teaching hospital was amplified by the polymerase chain reaction (PCR). The PCR products targeted -2100 nt regions of relatively low G + C content. Comparison of restriction enzyme digests of amplified DNA showed variation useful for strain differentiation. Twelve nonrelated HSV-1
were differentiated from one another. In contrast, specimens epidemioIogically related in an outbreak were indistinguishable from each other. Restriction endonuclease analysis of amplified HSV-1 sequences appears to be useful for molecular epidemiology and laboratory quality control to detect possible contamination by PCR products.
INTRODUCTION
and may thus be inconvenient to perform. In addition, the presence of hypervariable regions within the repeat regions of HSV DNA can lead to confusion in the interpretation of patterns (Davison and Wilkie, 1981). We report here the first primers suitable for amplification of large sections of HSV DNA. The DNA produced by the polymerase chain reaction (PCR) is sufficiently abundant to be detected nonisotopically. We show that analysis of restriction enzyme fragment length polymorphisms (RFLPs) within these amplified regions of HSV DNA is convenient and useful for comparing epidemiologically related HSV isolates.
Human herpes simplex virus types 1 and 2 (HSV-1 and HSV-2) are common infectious agents. Although HSV infections are usually associated with localized recurrent facial and genital blistering lesions, life-threatening encephalitis and generalized neonatal infection may occur. Molecular techniques have enabled epidemiologic studies of the transmission and reactivation of the virus. Variation among isolates of similar serotypes (type 1 or type 2) is readily detected by restriction enzyme mapping of the isolated DNA. In contrast, epidemiologically related isolates may have identical patterns (Whitley, 1985; Buchman et al., 1978). Large-scale tissue culture for sufficient viral DNA production or use of 32p labeling is required to enable restriction mapping
MATERIALS A N D M E T H O D S Viral Cultures
From the Department of Pathology (T.H.H., B.A., S.M., D.N.), Universityof Iowa College of Medicine; and Veterans Affairs Medical Center (T.H.H.), Iowa City, Iowa, USA. Address reprint requests to Dr. T.H. Haugen, Department of Pathology, University of Iowa College of Medicine/ML145, Iowa City, IA 52242, USA. Received 5 February 1993; revised and accepted 28 April 1993. © 1993Elsevier Science Publishing Co., Inc. 655 Avenue of the Americas, New York, NY 10010 0732-8893/93/$6.00
Viral cultures were performed in the Clinical Virology laboratory of the University of Iowa Hospitals and Clinics, Iowa City, Iowa. Specimen sources included swabs of the eye, throat, oral and digital lesions, and bronchial washings and sputum. The specimens were inoculated onto human lung fibroblasts (MRC-5 cells) cultured in Minimal Essential Medium with Earl's salts plus 2% fetal bovine serum,
130
T.H. Haugen et al.
fungizone, penicillin, and streptomycin. The isolates were identified as HSV serotype 1 by immunofluorescent staining. When cultures reached a cytopathic effect of -75%, samples were stored at - 70~C. Patient data have been reported, and restriction enzyme analysis of the entire HSV-1 32p-labeled DNA was performed as previously described (Perl et al., 1992).
Primers The HSV-1 DNA sequences used were obtained from Genbank (accession numbers X14112 and X14112-2) based partially on published work (McGeoch et al., 1988). Regional percent guanosine and cytosine content (G + C%) analysis was performed by a computer program written by one of the authors (T.H.H.), which is available upon request. PCR primers used are described in Table 1. The oligonucleotides were produced with an Applied Biosystems model 491 synthesizer and purified by Sephadex G-25 chromatography.
DNA Preparation and PCR Substrate DNA for amplification by PCR was prepared from either cultures of HSV or directly from thawed clinical specimens. HSV strains from stored cultures were thawed and inoculated into fresh MRC5 cells in shell vials. When -75% cytopathic effect was observed, individual shell vials were scraped and particulate material was harvested by centrifugation for 5 min in a table-top microfuge. The pellets were digested with 40 ~l of 400 ~g/ml proteinase K in 0.1% Tween-20, 10 mM Tris-HC1, 1 mM EDTA pH 7.5, at 56°C for 12-14 h. The protease was inactivated by heating to 95°C for 10 min. DNAs from stored, thawed bronchoscopy or lesional swab specimens were prepared in a similar manner except - 5 ~l of material pelleted in a table-top centrifuge was used. DNAs, 1 ~I from the above digestion mixtures, were amplified in a volume of 100 ~l by using the protocol supplied by the Perkin-Elmer Cetus "Gene Amp" Kit, except reaction mixtures contained 20%
glycerol. Addition of glycerol improved the yield of these PCR products. Some lots of glycerol, however, strongly inhibited PCR. Taq polymerase, 2.5 U, was added after an initial 5-min incubation at 94°C. Thermocycle conditions consisted of 35 cycles of 1 min at 94°C, 30 s at 65°C, and 2 min at 72°C. Amplified products were analyzed by electrophoresis in gels containing 2% agarose in TBE buffer (89 mM Tris, 89 mM boric acid, I mM EDTA). PCR products (2 ~1) were digested individually with MspI, HhaI, HaeIII, and BstUI. Restriction fragments from each digestion were resolved on TBE-10% acrylamide gels and visualized after silver staining (Frost et al., 1991).
RESULTS Amplification of HSV-1 D N A DNA with high G + C content may be difficult to amplify using PCR due to unusual thermal stability of the DNA. HSV-1 has a relatively high G + C content of 68.3% (McGeoch et al., 1988). The DNA sequence of HSV-1 was analyzed for regions of lower G + C content. As shown in Figure 1, several regions contain stretches of 2000 nt with as low as 62% G + C. Primers to be used for PCRs were designed that spanned two such areas without regard for potential or known protein-coding regions. No attempt was made to assure primers would anneal to HSV-2 DNA. The lengths of primers were 28-30 nt so that the annealing temperatures would be high. Note, however, that the regions selected do not include the terminal repeats. In no case did DNA from a cultured HSV-1 viral sample fail to yield an amplified product of the appropriate size (data not shown). HSV-2 samples, as expected, failed to be amplified by PCR with these primers. The amplified DNA was digested with restriction enzymes chosen to give as many digestion products as possible > 50 nt in size. As presented in Figure 2, the separated fragments from typical digests show several different restriction patterns. Among these 12 samples from epidemiologically un-
TABLE 1 Synthetic Nucleotides Used in This Study Primer
Sequence (5'-3')
Nucleotide Number a
PCR Product Sizeb
A1 A2
CACGCCGCTCGAGTGCGAGAGCAGCTFC CGTCCTCAACAGCCAGATCGCGGTGACC
11491-11508 13652-13625
2135
B1 B2
CAGAGCGCGACGGCGGACCTGGCGATCCAG GCCGCGCCCTCGAAGCCGGCCCTGCGTCTG
89977-90006 92160-92131
2184
aNucleotide numbers are derived from the published HSV-1 sequence (McGeochet al., 1988). bPCR, polymerasechain reaction.
HSV RFLPs
131
HSV-1 TRL
IRL IRs TRs Us
U~ "\A
"\
10KB
•
d+ (9 70 \ 60
:
:
:
:
B
E2
#/ "~ ~j/~4..~.,.,/ ~ J :
:
:
:
":
~6 45
:
a
,
J
/
Z 0 1 2 3
4 5 6
7 8
9 10 11 12 13 14 15 16 17 18 19 20
Differences i
FIGURE 1 HSV-1 DNA G + C content within 2000-nt blocks. The map of HSV is superimposed to indicate structural features. The location of the regions amplified by primer pairs A or B (see Table 1) are indicated by arrows. linked patients, most isolates gave a p r e d o m i n a n t fragment pattern. By digesting both PCR products with each of four different enzymes, however, every isolate was distinguished from all others.
Variation A m o n g HSV-1 Isolates To assess the degree of variation detected a m o n g these samples, the n u m b e r of restriction fragment differences were tabulated for each of the 66 pairs of samples. A n y b a n d not present in the corresponding pair was scored as one difference. The average number of differences was 10.0/pair. None of the pairs failed to be differentiated using four restriction enzymes. The tabulation of the differences shows a distribution (Figure 3) with a standard deviation of 4.4 differences/pair. A pair showing only one dif-
FIGURE 3 Summary of restriction fragment length differences observed between pairs of epidemiologically unrelated HSV-1 isolates. DNA from 12 cultured isolates (66 pairs) was amplified with primer sets A and B. The amplified DNAs were individually digested with BstUI, MspI, HhaI, and HaeIII, and analyzed by acrylamide gel electrophoresis. Each restriction fragment not shared was scored as one difference. The average total number of differences was 10.0 with a standard deviation of 4.4/pair. ference would lie outside of the range defined by 2 standard deviations from the mean. Therefore, if this sample is representative, >98% of any unrelated specimen pairs will s h o w at least one difference. All of these specimens had been s h o w n previously to be distinct by restriction digestion analysis of 32p_ labeled whole virus DNA (Perl et al., 1992). Although r a n d o m HSV-1 strains could be differentiated from each other, it w o u l d be expected that specimens epidemiologically related should give similar RFLP patterns. HSV-1 viral DNA amplified from cultures of the index patient's bronchoscopy specimen s h o w e d an identical pattern (Figure 4, lane O
0
o-1 2 3
o-1 2 3 4 5 6 4 5 6 7 8 9101112
FIGURE 2 Restriction fragment length analysis of 12 epidemiologically unrelated HSV-1 isolates. Amplified DNA produced by polymerase chain reaction with primer pair A was digested with BstU1 (lanes 1-12) and resolved by acrylamide gel electrophoresis followed by silver staining. DNA from a HpaII digest of pUC19 was used as a size standard (lane pUC).
FIGURE 4 Restriction fragment length analysis of epidemiologically related HSV-1 specimens. Polymerase chain reaction amplification, digestion, and electrophoresis was performed as in Figure 2. Lanes 1, 2, and 4 were amplified from cultures of specimens from a documented outbreak. Lane 3 is from an unrelated cultured specimen. DNA in lane 5 was amplified from an extract of the original bronchoscopy specimen that was cultured in lane 1. DNA in lane 6 was amplified from an extract of the original lesional swab that was cultured in lane 3. Lane pUC is a HpaII digest of pUC19.
132
1) to DNA amplified from cultures of specimens from two other patients (Figure 4, lanes 2 and 4) whose HSV-1 infections had been previously documented to have been transmitted from the index patient (Perl et al., 1992). These results with epidemiologically related specimens were the same as obtained by RFLP analysis of whole 32p-labeled HSV DNA and show that RFLP analysis of whole viral DNA or PCR products provides similar results. PCR can amplify DNA from specimens containing few organisms. This suggested that the original specimens may be directly used for amplification and RFLP analysis. HSV-1 viral DNA amplified directly from the index patient's bronchoscopy specimen (Figure 4, lane 5) showed an identical pattern to the virus later cultured from this specimen (Figure 4, lane 1). Direct amplification of DNA from another patient's lesion also showed an identical RFLP pattern to the DNA amplified from the cultured virus (Figure 4, lanes 3 and 6). These results show that culturing the virus is not necessary for RFLP analysis. DISCUSSION RFLP analysis is a common and powerful method for comparison of viruses or microorganisms. This technique has been shown to be useful for molecular epidemiology of HSV-1 (Whitley, 1985; Buchman et al., 1978). However, either 32p labeling of virus DNA or large-scale culture and preparation of viral DNA by using an ultracentrifuge, followed by extraction with phenol and precipitation with ethanol, is required for analysis. These methods are outside the abilities of many clinical laboratories. In contrast, the RFLP analysis of the PCR products generated from HSV-1 described in this report requires only the ability to amplify DNA and to perform gel electrophoresis. Use of toxic chemicals such as phenol to extract DNA or of radioactive precursors to label DNA is not necessary. Thus, RFLP analysis of PCR products has several advantages over previous techniques. Digestion of the two HSV-1 PCR products in this study with each of four restriction endonucleases produces a total of -100 restriction fragments. In a comparison of these restriction fragment lengths, all 66 unrelated isolate pairs showed differences. These data suggest that differences between unrelated isolates will not be detected <2% of the time. The average number of differences was 10.
T.H. Haugen et al.
Taq DNA polymerase, the enzyme used for amplification of DNA in this study, lacks a proofreading function. Random errors in nucleotide incorporation occur at a low rate and may be demonstrated by sequence analysis of molecular clones of amplified DNA or by denaturing gradient gel electrophoresis (Keohavong and Thilly, 1989) of uncloned amplified DNA. Direct sequence analysis of the uncloned amplified DNA will not show these errors because they are randomly distributed among the amplified molecules. Therefore, these errors do not affect RFLP analysis of amplified DNA because the molar prevalence of any single error is too low. HSV-1 in the clinical laboratory is routinely identiffed by culture, a highly sensitive technique. The cultured virus was therefore the source of DNA for amplification. PCR, however, has the potential for the primary laboratory diagnosis of HSV infection. We were able to amplify DNA directly from patient specimens. RFLP analysis of these amplified DNA showed similar patterns to the amplified DNA from cultured virus derived from the specimen. Therefore, in principle, viral culture would not be necessary for RFLP analysis. PCR has shown promise as a sensitive test for HSV in clinical specimens (Lynas et al., 1989; Hardy et al., 1990). A major concern about the usefulness of PCR as a diagnostic method is possible false-positive results caused by laboratory or other contamination. Since contaminants would be expected to be due to a discrete source, they should yield characteristic restriction enzyme fragment length patterns. Restriction fragment analysis of PCR products may therefore be an important laboratory quality control procedure. RFLP analysis of PCR products can demonstrate epidemiologically consistent differences in a high proportion of cytomegalovirus (Chou, 1990) and Chlamydia isolates (Rodrigues et al., 1991; Frost et al., 1991). It would seem likely that this technique could be extended to other clinically relevant organisms, especially those that are difficult to propagate rapidly.
The authors gratefully acknowledge the encouragement of S. Chou at the initial stage of this study and the critical reading of the manuscript by M.A Pfaller and C.T. Lutz. This work was supported in part by Department of Veterans Affairs research funds.
REFERENCES
Buchman TG, Roizman B, Adams G, Stover BH (1978) Restriction endonuclease fingerprinting of herpes sim-
plex DNA: a novel epidemiological tool applied to a nosocomial outbreak. J Infect Dis 138:488-498.
HSV RFLPs
Chou S (1990) Differentiation of cytomegalovirus strains by restriction analysis of DNA sequences amplified from clinical specimens. J Infect Dis 162:738-742. Davison AJ, Wilkie NM (1981) Nucleotide sequences of the joint between the l and s segments of herpes simplex virus types 1 and 2. J Gen Virol 55:315-331. Frost EH, Deslandes S, Veilleux S, Bourgaux-Ramoisy D (1991) Typing Chlamydia trachomatis by detection of restriction fragment length polymorphism in the gene encoding the major outer membrane protein. J Infect Dis 163:1103-1107. Hardy DA, Arvin AM, Yasukawa LL, Bronzan RN, Lewinsohn DM, Hensleigh PA, Prober CG (1990) Use of polymerase chain reaction for successful identification of asymptomatic genital infection with herpes simplex virus in pregnant women at delivery. ] Infect Dis 162:10311035. Keohavong P, Thilly WG (1989) Fidelity of DNA polymerases in DNA amplification. Proc Natl Acad Sci USA 86:9253-9257.
133
Lynas C, Cook SD, Laycock KA, Bradfield JWB, Maitland NJ (1989) Detection of latent virus mRNA in tissues using the polymerase chain reaction. J Pathol 157:285289. McGeoch DJ, Dalrymple MA, Davidson AJ, Dolan A, Frame MC, McNab D, Perry LJ, Scott JE, Taylor P (1988) The complete DNA sequence of the long unique region in the genome of herpes simplex virus type 1. J Gen Virol 69:1531-1574. Perl TM, Haugen TH, Pfaller MA, Hollis R, Lakeman AD, Whitley RJ, Nicholson D, Hunter GA, Wenzel RP (1992) Transmission of herpes simplex virus type 1 in an intensive care unit. Ann Intern Med 117:584-586. Rodrigues P, Vekris A, de Barbeyrac B, Dutilk B, Bonnet J, Bebear C (1991) Typing of Chlamydia trachomatis by restriction endonuclease analysis of the amplified major outer membrane protein gene. J Clin Microbiol 29:11321136. Whitley RJ (1985) Epidemiology of herpes simplex viruses. In The Herpesviruses, vol 3. Ed, B. Roizman. New York: Plenum, pp 1-44.
DIAGN MICROBIOLINFECTDIS 1993;17:129-133
VIROLOGY
Restriction Enzyme Fragment Length Polymorphisms of Amplified Herpes Simplex Virus Type-1 DNA Provide Epidemiologic Information Thomas H. Haugen, Beth Alden, Sandra Matthey, and Donald Nicholson
Human herpes simplex type 1 (HSV-1) DNA of isolates from patients in a large teaching hospital was amplified by the polymerase chain reaction (PCR). The PCR products targeted -2100 nt regions of relatively low G + C content. Comparison of restriction enzyme digests of amplified DNA showed variation useful for strain differentiation. Twelve nonrelated HSV-1
were differentiated from one another. In contrast, specimens epidemioIogically related in an outbreak were indistinguishable from each other. Restriction endonuclease analysis of amplified HSV-1 sequences appears to be useful for molecular epidemiology and laboratory quality control to detect possible contamination by PCR products.
INTRODUCTION
and may thus be inconvenient to perform. In addition, the presence of hypervariable regions within the repeat regions of HSV DNA can lead to confusion in the interpretation of patterns (Davison and Wilkie, 1981). We report here the first primers suitable for amplification of large sections of HSV DNA. The DNA produced by the polymerase chain reaction (PCR) is sufficiently abundant to be detected nonisotopically. We show that analysis of restriction enzyme fragment length polymorphisms (RFLPs) within these amplified regions of HSV DNA is convenient and useful for comparing epidemiologically related HSV isolates.
Human herpes simplex virus types 1 and 2 (HSV-1 and HSV-2) are common infectious agents. Although HSV infections are usually associated with localized recurrent facial and genital blistering lesions, life-threatening encephalitis and generalized neonatal infection may occur. Molecular techniques have enabled epidemiologic studies of the transmission and reactivation of the virus. Variation among isolates of similar serotypes (type 1 or type 2) is readily detected by restriction enzyme mapping of the isolated DNA. In contrast, epidemiologically related isolates may have identical patterns (Whitley, 1985; Buchman et al., 1978). Large-scale tissue culture for sufficient viral DNA production or use of 32p labeling is required to enable restriction mapping
MATERIALS A N D M E T H O D S Viral Cultures
From the Department of Pathology (T.H.H., B.A., S.M., D.N.), Universityof Iowa College of Medicine; and Veterans Affairs Medical Center (T.H.H.), Iowa City, Iowa, USA. Address reprint requests to Dr. T.H. Haugen, Department of Pathology, University of Iowa College of Medicine/ML145, Iowa City, IA 52242, USA. Received 5 February 1993; revised and accepted 28 April 1993. © 1993Elsevier Science Publishing Co., Inc. 655 Avenue of the Americas, New York, NY 10010 0732-8893/93/$6.00
Viral cultures were performed in the Clinical Virology laboratory of the University of Iowa Hospitals and Clinics, Iowa City, Iowa. Specimen sources included swabs of the eye, throat, oral and digital lesions, and bronchial washings and sputum. The specimens were inoculated onto human lung fibroblasts (MRC-5 cells) cultured in Minimal Essential Medium with Earl's salts plus 2% fetal bovine serum,
130
T.H. Haugen et al.
fungizone, penicillin, and streptomycin. The isolates were identified as HSV serotype 1 by immunofluorescent staining. When cultures reached a cytopathic effect of -75%, samples were stored at - 70~C. Patient data have been reported, and restriction enzyme analysis of the entire HSV-1 32p-labeled DNA was performed as previously described (Perl et al., 1992).
Primers The HSV-1 DNA sequences used were obtained from Genbank (accession numbers X14112 and X14112-2) based partially on published work (McGeoch et al., 1988). Regional percent guanosine and cytosine content (G + C%) analysis was performed by a computer program written by one of the authors (T.H.H.), which is available upon request. PCR primers used are described in Table 1. The oligonucleotides were produced with an Applied Biosystems model 491 synthesizer and purified by Sephadex G-25 chromatography.
DNA Preparation and PCR Substrate DNA for amplification by PCR was prepared from either cultures of HSV or directly from thawed clinical specimens. HSV strains from stored cultures were thawed and inoculated into fresh MRC5 cells in shell vials. When -75% cytopathic effect was observed, individual shell vials were scraped and particulate material was harvested by centrifugation for 5 min in a table-top microfuge. The pellets were digested with 40 ~l of 400 ~g/ml proteinase K in 0.1% Tween-20, 10 mM Tris-HC1, 1 mM EDTA pH 7.5, at 56°C for 12-14 h. The protease was inactivated by heating to 95°C for 10 min. DNAs from stored, thawed bronchoscopy or lesional swab specimens were prepared in a similar manner except - 5 ~l of material pelleted in a table-top centrifuge was used. DNAs, 1 ~I from the above digestion mixtures, were amplified in a volume of 100 ~l by using the protocol supplied by the Perkin-Elmer Cetus "Gene Amp" Kit, except reaction mixtures contained 20%
glycerol. Addition of glycerol improved the yield of these PCR products. Some lots of glycerol, however, strongly inhibited PCR. Taq polymerase, 2.5 U, was added after an initial 5-min incubation at 94°C. Thermocycle conditions consisted of 35 cycles of 1 min at 94°C, 30 s at 65°C, and 2 min at 72°C. Amplified products were analyzed by electrophoresis in gels containing 2% agarose in TBE buffer (89 mM Tris, 89 mM boric acid, I mM EDTA). PCR products (2 ~1) were digested individually with MspI, HhaI, HaeIII, and BstUI. Restriction fragments from each digestion were resolved on TBE-10% acrylamide gels and visualized after silver staining (Frost et al., 1991).
RESULTS Amplification of HSV-1 D N A DNA with high G + C content may be difficult to amplify using PCR due to unusual thermal stability of the DNA. HSV-1 has a relatively high G + C content of 68.3% (McGeoch et al., 1988). The DNA sequence of HSV-1 was analyzed for regions of lower G + C content. As shown in Figure 1, several regions contain stretches of 2000 nt with as low as 62% G + C. Primers to be used for PCRs were designed that spanned two such areas without regard for potential or known protein-coding regions. No attempt was made to assure primers would anneal to HSV-2 DNA. The lengths of primers were 28-30 nt so that the annealing temperatures would be high. Note, however, that the regions selected do not include the terminal repeats. In no case did DNA from a cultured HSV-1 viral sample fail to yield an amplified product of the appropriate size (data not shown). HSV-2 samples, as expected, failed to be amplified by PCR with these primers. The amplified DNA was digested with restriction enzymes chosen to give as many digestion products as possible > 50 nt in size. As presented in Figure 2, the separated fragments from typical digests show several different restriction patterns. Among these 12 samples from epidemiologically un-
TABLE 1 Synthetic Nucleotides Used in This Study Primer
Sequence (5'-3')
Nucleotide Number a
PCR Product Sizeb
A1 A2
CACGCCGCTCGAGTGCGAGAGCAGCTFC CGTCCTCAACAGCCAGATCGCGGTGACC
11491-11508 13652-13625
2135
B1 B2
CAGAGCGCGACGGCGGACCTGGCGATCCAG GCCGCGCCCTCGAAGCCGGCCCTGCGTCTG
89977-90006 92160-92131
2184
aNucleotide numbers are derived from the published HSV-1 sequence (McGeochet al., 1988). bPCR, polymerasechain reaction.
HSV RFLPs
131
HSV-1 TRL
IRL IRs TRs Us
U~ "\A
"\
10KB
•
d+ (9 70 \ 60
:
:
:
:
B
E2
#/ "~ ~j/~4..~.,.,/ ~ J :
:
:
:
":
~6 45
:
a
,
J
/
Z 0 1 2 3
4 5 6
7 8
9 10 11 12 13 14 15 16 17 18 19 20
Differences i
FIGURE 1 HSV-1 DNA G + C content within 2000-nt blocks. The map of HSV is superimposed to indicate structural features. The location of the regions amplified by primer pairs A or B (see Table 1) are indicated by arrows. linked patients, most isolates gave a p r e d o m i n a n t fragment pattern. By digesting both PCR products with each of four different enzymes, however, every isolate was distinguished from all others.
Variation A m o n g HSV-1 Isolates To assess the degree of variation detected a m o n g these samples, the n u m b e r of restriction fragment differences were tabulated for each of the 66 pairs of samples. A n y b a n d not present in the corresponding pair was scored as one difference. The average number of differences was 10.0/pair. None of the pairs failed to be differentiated using four restriction enzymes. The tabulation of the differences shows a distribution (Figure 3) with a standard deviation of 4.4 differences/pair. A pair showing only one dif-
FIGURE 3 Summary of restriction fragment length differences observed between pairs of epidemiologically unrelated HSV-1 isolates. DNA from 12 cultured isolates (66 pairs) was amplified with primer sets A and B. The amplified DNAs were individually digested with BstUI, MspI, HhaI, and HaeIII, and analyzed by acrylamide gel electrophoresis. Each restriction fragment not shared was scored as one difference. The average total number of differences was 10.0 with a standard deviation of 4.4/pair. ference would lie outside of the range defined by 2 standard deviations from the mean. Therefore, if this sample is representative, >98% of any unrelated specimen pairs will s h o w at least one difference. All of these specimens had been s h o w n previously to be distinct by restriction digestion analysis of 32p_ labeled whole virus DNA (Perl et al., 1992). Although r a n d o m HSV-1 strains could be differentiated from each other, it w o u l d be expected that specimens epidemiologically related should give similar RFLP patterns. HSV-1 viral DNA amplified from cultures of the index patient's bronchoscopy specimen s h o w e d an identical pattern (Figure 4, lane O
0
o-1 2 3
o-1 2 3 4 5 6 4 5 6 7 8 9101112
FIGURE 2 Restriction fragment length analysis of 12 epidemiologically unrelated HSV-1 isolates. Amplified DNA produced by polymerase chain reaction with primer pair A was digested with BstU1 (lanes 1-12) and resolved by acrylamide gel electrophoresis followed by silver staining. DNA from a HpaII digest of pUC19 was used as a size standard (lane pUC).
FIGURE 4 Restriction fragment length analysis of epidemiologically related HSV-1 specimens. Polymerase chain reaction amplification, digestion, and electrophoresis was performed as in Figure 2. Lanes 1, 2, and 4 were amplified from cultures of specimens from a documented outbreak. Lane 3 is from an unrelated cultured specimen. DNA in lane 5 was amplified from an extract of the original bronchoscopy specimen that was cultured in lane 1. DNA in lane 6 was amplified from an extract of the original lesional swab that was cultured in lane 3. Lane pUC is a HpaII digest of pUC19.
132
1) to DNA amplified from cultures of specimens from two other patients (Figure 4, lanes 2 and 4) whose HSV-1 infections had been previously documented to have been transmitted from the index patient (Perl et al., 1992). These results with epidemiologically related specimens were the same as obtained by RFLP analysis of whole 32p-labeled HSV DNA and show that RFLP analysis of whole viral DNA or PCR products provides similar results. PCR can amplify DNA from specimens containing few organisms. This suggested that the original specimens may be directly used for amplification and RFLP analysis. HSV-1 viral DNA amplified directly from the index patient's bronchoscopy specimen (Figure 4, lane 5) showed an identical pattern to the virus later cultured from this specimen (Figure 4, lane 1). Direct amplification of DNA from another patient's lesion also showed an identical RFLP pattern to the DNA amplified from the cultured virus (Figure 4, lanes 3 and 6). These results show that culturing the virus is not necessary for RFLP analysis. DISCUSSION RFLP analysis is a common and powerful method for comparison of viruses or microorganisms. This technique has been shown to be useful for molecular epidemiology of HSV-1 (Whitley, 1985; Buchman et al., 1978). However, either 32p labeling of virus DNA or large-scale culture and preparation of viral DNA by using an ultracentrifuge, followed by extraction with phenol and precipitation with ethanol, is required for analysis. These methods are outside the abilities of many clinical laboratories. In contrast, the RFLP analysis of the PCR products generated from HSV-1 described in this report requires only the ability to amplify DNA and to perform gel electrophoresis. Use of toxic chemicals such as phenol to extract DNA or of radioactive precursors to label DNA is not necessary. Thus, RFLP analysis of PCR products has several advantages over previous techniques. Digestion of the two HSV-1 PCR products in this study with each of four restriction endonucleases produces a total of -100 restriction fragments. In a comparison of these restriction fragment lengths, all 66 unrelated isolate pairs showed differences. These data suggest that differences between unrelated isolates will not be detected <2% of the time. The average number of differences was 10.
T.H. Haugen et al.
Taq DNA polymerase, the enzyme used for amplification of DNA in this study, lacks a proofreading function. Random errors in nucleotide incorporation occur at a low rate and may be demonstrated by sequence analysis of molecular clones of amplified DNA or by denaturing gradient gel electrophoresis (Keohavong and Thilly, 1989) of uncloned amplified DNA. Direct sequence analysis of the uncloned amplified DNA will not show these errors because they are randomly distributed among the amplified molecules. Therefore, these errors do not affect RFLP analysis of amplified DNA because the molar prevalence of any single error is too low. HSV-1 in the clinical laboratory is routinely identiffed by culture, a highly sensitive technique. The cultured virus was therefore the source of DNA for amplification. PCR, however, has the potential for the primary laboratory diagnosis of HSV infection. We were able to amplify DNA directly from patient specimens. RFLP analysis of these amplified DNA showed similar patterns to the amplified DNA from cultured virus derived from the specimen. Therefore, in principle, viral culture would not be necessary for RFLP analysis. PCR has shown promise as a sensitive test for HSV in clinical specimens (Lynas et al., 1989; Hardy et al., 1990). A major concern about the usefulness of PCR as a diagnostic method is possible false-positive results caused by laboratory or other contamination. Since contaminants would be expected to be due to a discrete source, they should yield characteristic restriction enzyme fragment length patterns. Restriction fragment analysis of PCR products may therefore be an important laboratory quality control procedure. RFLP analysis of PCR products can demonstrate epidemiologically consistent differences in a high proportion of cytomegalovirus (Chou, 1990) and Chlamydia isolates (Rodrigues et al., 1991; Frost et al., 1991). It would seem likely that this technique could be extended to other clinically relevant organisms, especially those that are difficult to propagate rapidly.
The authors gratefully acknowledge the encouragement of S. Chou at the initial stage of this study and the critical reading of the manuscript by M.A Pfaller and C.T. Lutz. This work was supported in part by Department of Veterans Affairs research funds.
REFERENCES
Buchman TG, Roizman B, Adams G, Stover BH (1978) Restriction endonuclease fingerprinting of herpes sim-
plex DNA: a novel epidemiological tool applied to a nosocomial outbreak. J Infect Dis 138:488-498.
HSV RFLPs
Chou S (1990) Differentiation of cytomegalovirus strains by restriction analysis of DNA sequences amplified from clinical specimens. J Infect Dis 162:738-742. Davison AJ, Wilkie NM (1981) Nucleotide sequences of the joint between the l and s segments of herpes simplex virus types 1 and 2. J Gen Virol 55:315-331. Frost EH, Deslandes S, Veilleux S, Bourgaux-Ramoisy D (1991) Typing Chlamydia trachomatis by detection of restriction fragment length polymorphism in the gene encoding the major outer membrane protein. J Infect Dis 163:1103-1107. Hardy DA, Arvin AM, Yasukawa LL, Bronzan RN, Lewinsohn DM, Hensleigh PA, Prober CG (1990) Use of polymerase chain reaction for successful identification of asymptomatic genital infection with herpes simplex virus in pregnant women at delivery. ] Infect Dis 162:10311035. Keohavong P, Thilly WG (1989) Fidelity of DNA polymerases in DNA amplification. Proc Natl Acad Sci USA 86:9253-9257.
133
Lynas C, Cook SD, Laycock KA, Bradfield JWB, Maitland NJ (1989) Detection of latent virus mRNA in tissues using the polymerase chain reaction. J Pathol 157:285289. McGeoch DJ, Dalrymple MA, Davidson AJ, Dolan A, Frame MC, McNab D, Perry LJ, Scott JE, Taylor P (1988) The complete DNA sequence of the long unique region in the genome of herpes simplex virus type 1. J Gen Virol 69:1531-1574. Perl TM, Haugen TH, Pfaller MA, Hollis R, Lakeman AD, Whitley RJ, Nicholson D, Hunter GA, Wenzel RP (1992) Transmission of herpes simplex virus type 1 in an intensive care unit. Ann Intern Med 117:584-586. Rodrigues P, Vekris A, de Barbeyrac B, Dutilk B, Bonnet J, Bebear C (1991) Typing of Chlamydia trachomatis by restriction endonuclease analysis of the amplified major outer membrane protein gene. J Clin Microbiol 29:11321136. Whitley RJ (1985) Epidemiology of herpes simplex viruses. In The Herpesviruses, vol 3. Ed, B. Roizman. New York: Plenum, pp 1-44.